Computer Science Learning Curve Survey 2024 Report
In 2024, JetBrains Academy surveyed 23,991 respondents worldwide, including university students, online learners, self-taught enthusiasts, coding boot camp graduates, professionals, and career switchers.
Based on their inspiring insights, this report explores the current trends in computer science education, from formats and tools to motivations, career goals, and challenges.
Whether you're an educator, researcher, learner, curious professional, or supportive parent, dive in! Share your thoughts and connect with the CS learning community using #JetBrainsAcademySurvey24.
This is a public report; its contents may be used only for non-commercial purposes. Get the full details here.
Share:
Formal Education
Have you studied computer science in any format over the past 12 months?
Yes, self-education
Yes, at a formal educational institution
Just over half of computer science learners study at formal educational institutions, with 54% of formal learners broadening their knowledge through further self-education.
of those who have completed formal education hold a bachelor's degree or higher.
Highest level of formal education completed
Formal educational institution currently attended
Degree currently pursued
Major field of study (current or past)
Computer science
Software engineering
Other engineering
Art / Humanities
Economics
Mathematics
Biology / Chemistry
Social sciences
Physics
Other
Career
of all computer science learners have paid IT work experience, and for 89% of them, this is their primary income source. Most of these respondents work in software engineering roles (76%), with 35% holding mid-level positions.
Current job role100+
This question was only shown to those who responded that they rely on work in computer science as their primary income source.
The tech industry remains predominantly male across most roles, with significantly lower representation for women and non-binary individuals. Core technical roles and leadership positions (team leads and executives) have the least gender diversity, with 88%–94% male representation.
However, some roles show relatively higher female representation compared to the industry average: UX/UI designers (16%), QA/Testers, business analysts (14%), instructors (13%), and product/marketing managers (12%). Non-binary representation remains limited across all roles, with developer advocates seeing the highest rate at 6%.
Employment status
Seniority level
Salary (annual net in USD, excluding any bonuses)
Do you have previous work experience outside of computer science/IT?
Yes, I worked/studied in another field before switching to computer science/IT
No, this is the only field I’ve ever worked in
Younger respondents aged 18–29 are more likely to go directly into a tech career, with only 9% of those aged 18–20 and 24% of those aged 21–29 having prior experience in another field. However, career switching becomes more common with age, with 50% of respondents aged 30–39 and 59% of those aged 60+ reporting previous careers outside of tech.
There are also clear regional differences in career trajectories. In India and China, non-career-switchers form the majority, reflecting a strong trend of direct entry into computer science. In contrast, Argentina and Brazil show more diverse pathways, with career switchers either outnumbering or nearly matching non-switchers. In regions like Europe, Southeast Asia, and North America, career switchers make up about one-third, reflecting a more conventional entry pattern.
Do you have previous work experience outside of computer science/IT? (by region)
India
China
Germany
Türkiye
Middle East, Africa, Central Asia
Other Southeast Asia and Oceania
South Korea
Rest of Europe
France
Canada
Previous professional sphere
Respondents answered this question with open-text answers. ChatGPT was used to automate the analysis and sorting of the responses into thematic clusters.
This question was only shown to respondents who said they had worked or studied in another field before switching to computer science/IT.
Engineering and technical fields take the top spot among those transitioning to computer science, followed by finance and business management. Education, healthcare, and creative arts also rank prominently, showcasing the diverse professional backgrounds entering the field.
Reasons for choosing a career in tech100+
Motivation for learning new computer science topics100+
While a strong passion for computer science drives most career transitions, nearly half of respondents highlight their love for problem-solving and process automation as key motivators. Interestingly, salary and remote work opportunities rank slightly lower than creative ambitions, such as building games or websites, revealing that the field attracts those driven by aspirations as much as by practical benefits.
Motivation for learning new computer science topics (by region)
| I don’t want to learn new computer science topics | Other | To complete a specific task | To migrate to another technology | Out of interest | To find a new job or switch roles | To keep up with the latest trends | To work on personal projects | To grow in my current role | |
|---|---|---|---|---|---|---|---|---|---|
| <1% | 2% | 18% | 16% | 43% | 52% | 49% | 56% | 68% | Eastern Europe, Balkans, and the Caucasus |
| <1% | 1% | 13% | 11% | 49% | 49% | 40% | 49% | 67% | South Korea |
| <1% | 2% | 26% | 21% | 47% | 47% | 51% | 56% | 67% | Other Southeast Asia and Oceania |
| <1% | 2% | 27% | 19% | 79% | 34% | 48% | 60% | 66% | Germany |
| – | 3% | 21% | 17% | 67% | 44% | 47% | 55% | 64% | Benelux and Northern Europe |
| 1% | 2% | 17% | 17% | 45% | 50% | 55% | 59% | 64% | India |
| <1% | 1% | 22% | 26% | 23% | 45% | 55% | 49% | 64% | Nigeria |
| <1% | 2% | 20% | 18% | 51% | 46% | 47% | 58% | 62% | Rest of Europe |
| – | <1% | 23% | 17% | 67% | 43% | 47% | 44% | 62% | China |
| – | 2% | 21% | 14% | 62% | 48% | 44% | 58% | 61% | United Kingdom |
| 1% | 2% | 22% | 16% | 58% | 54% | 45% | 65% | 61% | United States |
| 1% | 2% | 19% | 21% | 38% | 44% | 48% | 54% | 60% | Middle East, Africa, Central Asia |
| – | 3% | 13% | 18% | 58% | 50% | 54% | 51% | 60% | Spain |
| 1% | 1% | 20% | 22% | 45% | 41% | 46% | 51% | 56% | Türkiye |
| <1% | 2% | 25% | 13% | 56% | 59% | 45% | 62% | 56% | Canada |
| 2% | 1% | 15% | 19% | 42% | 41% | 28% | 39% | 55% | Russian Federation and Belarus |
| – | 3% | 16% | 21% | 52% | 64% | 42% | 57% | 54% | Brazil |
| 1% | 1% | 24% | 23% | 73% | 38% | 39% | 58% | 54% | France |
| 9% | 1% | 10% | 18% | 49% | 63% | 46% | 56% | 54% | Mexico |
| <1% | 2% | 11% | 19% | 41% | 60% | 51% | 57% | 52% | Central and South America |
| 4% | <1% | 14% | 19% | 43% | 40% | 31% | 38% | 50% | Ukraine |
| 3% | 1% | 12% | 13% | 58% | 34% | 42% | 31% | 48% | Japan |
| 1% | 2% | 9% | 17% | 52% | 63% | 44% | 47% | 42% | Argentina |
Developed regions, such as Western Europe and North America, show stability, with learners focusing on personal interests and innovative personal projects. In contrast, learners in Latin America are motivated by the opportunity to switch jobs, which reflects fluid job markets but also a lesser emphasis on immediate practical skills. Asia shows a spectrum of motivations. South Korea aligns with career-driven growth, while Japan reports low engagement across learning dimensions, indicating a potential need for policy and cultural shifts. In India and Southeast Asia, learners are motivated to keep up with trends, which reflects the dynamism of their growing tech ecosystems.
Desired job role100+
This question was only shown to respondents who indicated “finding a new job” or “switching roles” as one of their motivations to learn computer science topics.
Developer is the top career choice in IT, most likely a reflection of the role’s versatility, high demand, and broad applicability across industries, making it an optimal choice for career transitions, especially for individuals new to the field. Significant numbers are also branching into data-focused careers or DevOps, showcasing the growing appeal of specialized fields. On the contrary, QA roles, though good for entry, lack popularity and long-term prospects, making them less aspirational for career transitions.
of respondents report they have, at one point, searched for work in computer science/IT.
Important factors when searching for a job in computer science/IT
| Not important | Fairly unimportant | Fairly important | Extremely important | |
|---|---|---|---|---|
| 1% | 6% | 35% | 58% | Work experience |
| 1% | 13% | 51% | 35% | Familiarity with the latest technologies |
| 2% | 16% | 51% | 32% | Soft skills |
| 4% | 17% | 47% | 31% | Internships and co-op programs |
| 6% | 26% | 44% | 25% | Connections and networking |
| 5% | 23% | 48% | 24% | Pet projects |
| 7% | 26% | 49% | 18% | University diplomas |
| 6% | 31% | 47% | 16% | Peer references |
| 9% | 31% | 46% | 14% | Industry certificates |
| 11% | 35% | 42% | 12% | Course completion certificates |
Work experience and up-to-date tech knowledge are reportedly key to landing a job, but soft skills are equally valued, with 83% of learners marking them as important. Networking is another crucial factor – 25% consider it critical, and 44% actively use their connections for career opportunities. This underscores the need for strong interpersonal skills and professional networks in the tech sector.
Learning Topics
Computer science areas studied in the past three years100+
Along with programming languages, algorithms, and data structures, databases are a popular choice for learners. AI and machine learning remain popular fields, with 33% and 34% of learners exploring them, respectively.
Proficiency in computer science areas studied
| Novice / Exploratory | Beginner | Intermediate | Advanced | Expert | |
|---|---|---|---|---|---|
| 4% | 25% | 44% | 23% | 5% | Software engineering |
| 6% | 28% | 41% | 21% | 5% | Web development |
| 8% | 29% | 40% | 17% | 5% | Product management |
| 4% | 23% | 47% | 22% | 4% | Programming languages |
| 10% | 33% | 37% | 16% | 4% | Human-computer interaction (HCI) |
| 9% | 33% | 38% | 16% | 4% | Project management |
| 9% | 37% | 37% | 14% | 3% | Testing |
| 10% | 37% | 35% | 15% | 3% | Data analysis |
| 15% | 42% | 30% | 10% | 3% | Natural language processing (NLP) |
| 16% | 40% | 29% | 11% | 3% | Computer vision |
| 7% | 32% | 41% | 16% | 3% | Databases |
| 9% | 36% | 39% | 13% | 3% | Computer networks |
| 7% | 35% | 38% | 16% | 3% | Operating systems |
| 11% | 40% | 34% | 12% | 3% | Cybersecurity |
| 6% | 31% | 46% | 15% | 2% | Algorithms and data structures |
| 17% | 43% | 27% | 10% | 2% | Artificial intelligence |
| 18% | 43% | 27% | 10% | 2% | Machine learning |
| 16% | 39% | 30% | 13% | 2% | Computer graphics |
Women tend to rate their technical skills lower, yet they demonstrate a strong drive for growth, with 8% more female learners transitioning into computer science from other fields compared to their male counterparts.
Topics learners want to explore in next course
Respondents answered this question with open-text answers. ChatGPT was used to automate the analysis and sorting of the responses into thematic clusters.
Learning Formats, Practices, and Resources
Experience with educational formats100+
The data shows a continued demand for traditional, in-person, hands-on learning environments like university education, workshops, and mentorship programs. However, satisfaction with these formats varies widely across age groups and regions, reflecting inconsistent effectiveness.
Rating of experience with educational formats
| Poor | Needs Improvement | Satisfactory | Very good | Excellent | |
|---|---|---|---|---|---|
| 2% | 5% | 18% | 32% | 43% | Internships |
| 1% | 7% | 22% | 36% | 34% | Mentorship programs and tutoring |
| 2% | 4% | 22% | 42% | 30% | Paid online courses (MOOCs) or code schools |
| 2% | 9% | 23% | 38% | 29% | Professional training provided by an employer |
| 1% | 5% | 25% | 40% | 29% | Self-paced online tutorials |
| 1% | 7% | 28% | 36% | 28% | Codecamps, user groups, meetups |
| 3% | 9% | 29% | 33% | 26% | Vocational programs |
| 4% | 8% | 26% | 37% | 25% | Outsourced professional training, paid for by an employer |
| 2% | 7% | 26% | 40% | 25% | Coding bootcamp sessions |
| 2% | 9% | 31% | 34% | 24% | Exchange programs |
| 1% | 8% | 31% | 38% | 21% | Free online courses (MOOCs) or code schools |
| 3% | 11% | 31% | 36% | 20% | Online university programs |
| 6% | 14% | 31% | 30% | 19% | University, college, school education |
| 2% | 11% | 32% | 36% | 19% | Offline courses, code schools |
| 2% | 10% | 33% | 36% | 19% | Workshops and seminars |
University, college, and school education, as well as self-paced online tutorials, are top answers for all respondents. The rest depends on the specific age group and career needs. Workshops and seminars are most popular among the 50–59 age group, with 17% of such learners having experience with them and about one quarter rating their experience as excellent. Mentorship programs are highly rated by respondents aged 21–29, with 36% of them rating it as excellent, but the satisfaction with this experience declines as age increases. Satisfaction with employer-provided training peaks among respondents aged 18–20, with 41% of learners marking it as excellent. Paid online courses and coding bootcamps appeal most to younger and mid-career individuals.
Familiarity with MOOCs and code schools
| Never heard of it | Aware of it, but never tried it | Tried it, but don’t use it anymore | Currently use it | |
|---|---|---|---|---|
| 18% | 23% | 29% | 29% | Udemy |
| 18% | 27% | 35% | 20% | Coursera |
| 29% | 41% | 15% | 16% | JetBrains Academy |
| 38% | 26% | 25% | 10% | edX |
| 26% | 35% | 29% | 10% | Codecademy |
| 35% | 36% | 20% | 10% | LinkedIn Learning |
| 28% | 33% | 30% | 9% | Khan Academy |
| 52% | 28% | 10% | 9% | Canvas |
| 55% | 28% | 12% | 5% | DataCamp |
| 48% | 32% | 16% | 4% | Udacity |
| 67% | 18% | 11% | 4% | Pluralsight |
| 79% | 13% | 5% | 3% | Stepik |
| 85% | 8% | 4% | 3% | SWAYAM |
| 84% | 11% | 4% | 2% | JavaRush |
| 70% | 22% | 6% | 2% | The Open University |
| 78% | 16% | 5% | 1% | FutureLearn |
| 84% | 12% | 3% | 1% | Egghead |
| 90% | 7% | 2% | 1% | XuetangX |
| 92% | 6% | 2% | 1% | MiríadaX |
| 89% | 8% | 2% | 1% | Cognitive Class |
| 87% | 9% | 3% | 1% | Platzi |
Did you know?
JetBrains Academy users are 24% more likely to rate their experience with paid online courses (MOOCs) or code schools as “Excellent”. Discover your learning options with JetBrains Academy.
Practices for mastering computer science topics100+
Learners exploring computer science prioritize hands-on and visual learning, with coding platforms, video tutorials, and documentation leading the way. However, the recent stats on AI chatbot usage and participation in coding contests imply a shift toward interactive and dynamic approaches to problem-solving and skill-building.
This blend of traditional and modern resources suggests that learners value both structured guidance and opportunities for creative experimentation.
Competitive coding experience
Extensive experience: I regularly compete or have competed a lot in the past
Moderate experience: I’ve participated in a few contests
No interest: I don’t have any experience in it, nor do I want to compete
No experience: I’m new to competitive coding but interested in it
The majority of respondents are new to competitive coding but interested in it, while 30% have some experience and have participated in a few contests or used to compete regularly in the past.
Preferred resources and communities for learning computer science100+
Peer interaction is a key component of CS learning. About one-third of respondents value hackathons and open-source contributions, while a quarter prefer engaging with coding communities for learning. While platforms and tutorials dominate, collaborative and competitive activities inspire deeper engagement.
100+
Learners of all ages rely on various resources for help. Google is the top choice for all ages, while AI assistants like ChatGPT are especially popular among younger users, with two-thirds of those under 29 using them. Younger learners also tend to seek help from friends and educators, while those in their 30s and 40s turn to colleagues. YouTube is widely used across all ages, while older learners prefer textbooks and platforms like Medium. Overall, younger generations balance AI, peer support, and educational media, while older groups favor professional networks, structured articles, and textbooks.
100+
| 18–20 | 21–29 | 30–39 | 40–49 | 50–59 | 60 or older | |
|---|---|---|---|---|---|---|
| 70% | 76% | 77% | 75% | 68% | 68% | |
| 66% | 67% | 55% | 46% | 38% | 35% | An AI-based assistant (ChatGPT or other) |
| 58% | 48% | 31% | 22% | 13% | 9% | Friends and classmates |
| 56% | 65% | 64% | 52% | 37% | 33% | Stack Overflow |
| 53% | 53% | 50% | 50% | 43% | 36% | YouTube |
| 47% | 32% | 21% | 20% | 16% | 8% | An educator / teacher / tutor |
| 23% | 23% | 27% | 32% | 31% | 34% | Textbooks |
| 20% | 25% | 26% | 27% | 18% | 25% | Online tech media (e.g. Medium) |
| 19% | 19% | 18% | 16% | 24% | 14% | People on social media |
| 17% | 31% | 36% | 33% | 29% | 19% | Colleagues |
| 3% | 3% | 3% | 4% | 2% | 9% | Other |
of respondents report using AI assistants in their everyday life.
Most popular AI assistants in use100+
All answers with less than a 1% share have been merged into “Other”.
Most popular AI assistant functionality for learning computer science100+
AI helps learners overcome language barriers. Given that English is the dominant language for most computer science resources, regions with diverse languages or primarily non-English-speaking populations depend more on translation and pronunciation functionality.
The highest reliance on these features is seen in Northern Eurasia (44%) and Turkey (45%), followed closely by South and East Asia, Latin America, and Southeast Asia and Oceania (in these regions the share of such functionality usage varies from 40% to 44%).
In contrast, predominantly English-speaking countries like the United Kingdom, Canada, and the United States exhibit much lower levels of usage (13%–19%), reflecting fewer language-related challenges for learners.
Choosing a Course and Investment
The most important aspects for learners choosing a course are hands-on projects and exercises for practical experience, access to resources and materials, affordable prices, and the instructor’s industry background.
Course design and content
| Of little importance | Fairly important | Very important | |
|---|---|---|---|
| 2% | 22% | 76% | Hands-on projects and exercises for practical experience |
| 3% | 31% | 66% | Structured curriculum with progressive topics |
| 3% | 32% | 65% | Clear learning objectives for students |
| 7% | 32% | 60% | Real-world relevance |
| 6% | 38% | 55% | Responsiveness to changing industry standards |
| 7% | 41% | 52% | Simplification of complex concepts for all levels |
| 9% | 44% | 47% | Responsiveness to student feedback |
| 17% | 40% | 42% | Career development guidance |
| 20% | 43% | 38% | Ethical considerations regarding responsible use of technology |
| 31% | 46% | 24% | Peer collaboration |
| 38% | 41% | 21% | Gamification (quizzes, badges, etc.) |
Student support and flexibility
| Of little importance | Fairly important | Very important | |
|---|---|---|---|
| 2% | 25% | 74% | Access to resources and materials |
| 6% | 38% | 56% | Time flexibility |
| 10% | 35% | 54% | Remote studying options |
| 6% | 45% | 49% | Regular feedback and assessments |
| 12% | 44% | 44% | Supportive community and networking |
| 14% | 48% | 38% | Technical support services |
| 21% | 41% | 38% | Accessible study place |
| 24% | 40% | 36% | Offline studying options |
| 23% | 42% | 35% | Inclusive environment |
| 19% | 45% | 35% | Environmental accessibility |
| 53% | 33% | 14% | Daycare provision |
Female learners prioritize flexibility and support in education more than male learners do. Differences include a higher emphasis on time flexibility (64% for females vs. 54% for males), remote study options (63% vs. 53%), and technical support (50% vs. 36%). Additionally, 49% of female learners value accessible study spaces, compared to 36% of males.
Affordability
| Of little importance | Fairly important | Very important | |
|---|---|---|---|
| 3% | 32% | 66% | Affordable price |
| 17% | 48% | 35% | A customizable tariff structure allowing payment for individual components |
| 18% | 49% | 32% | Business-to-business (B2B) options available for convenient cost coverage by my employer |
Certification and credentials
| Of little importance | Fairly important | Very important | |
|---|---|---|---|
| 19% | 37% | 44% | University diploma of higher education |
| 16% | 41% | 43% | Industry certification |
| 21% | 40% | 39% | Certification or credentials upon course completion |
Although a university diploma was the top selection, all listed certification options are valuable for the general audience, validating acquired skills and knowledge.
Instructors’ qualifications and personality
| Of little importance | Fairly important | Very important | |
|---|---|---|---|
| 8% | 37% | 56% | Industry background |
| 15% | 46% | 39% | Empathy |
| 20% | 46% | 34% | Сharisma |
| 29% | 41% | 30% | Academic or university background |
Monthly spending on online education
Less than USD 25
USD 25–50
USD 51–100
USD 101–200
More than USD 200
I don’t spend money on online education
I’d prefer not to say
About three-quarters of respondents pay for online education. When it comes to current courses, high-quality and well-structured content, hands-on practice, and flexible formats are their three main reasons for opting for paid courses. When asked what would motivate them to pay for courses (or any other type of learning materials) in the future, respondents emphasized relevance to work/studies, personal interest, specialized content, and certification.
Reasons for paying for current courses100+
Note: Respondents answered this question with open-text answers. ChatGPT was used to automate the analysis and sorting of the responses into thematic clusters.
Reasons for paying for future courses
Note: Respondents answered this question with open-text answers. ChatGPT was used to automate the analysis and sorting of the responses into thematic clusters.
Learning Challenges
of computer science learners have quit a course, with the most common reasons cited as unengaging content, time constraints, and lack of practical exercises. Self-paced online tutorials and free MOOCs are the most commonly abandoned, pointing to challenges in staying motivated in less-structured learning formats.
Reasons for quitting a course or learning program100+
Type of course most recently quit
All answers with less than a 1% share have been merged into “Other”.
The most challenging aspects of studying computer science100+
Learners often struggle with practical hurdles like debugging and choosing the right resources, as well as emotional barriers like imposter syndrome and isolation. These insights highlight the dual need for clear guidance and supportive learning environments to help students thrive.
Methods of overcoming frustration
Note: Respondents answered this question with open-text answers. ChatGPT was used to automate the analysis and sorting of the responses into thematic clusters.
Our respondents’ most effective strategies for overcoming frustration include taking breaks and engaging in physical activities, as well as setting goals and reminding oneself of one’s initial motivations. Self-reflection and adjusting one's mindset also emerge as key approaches, helping individuals navigate challenges with a more adaptable and positive outlook. These methods assist learners in resetting, regaining focus, and recharging. However, 18% of respondents are still searching for effective solutions, highlighting the lack of universal solutions for managing frustration.
How learners keep themselves productive100+
Globally, breaking tasks down into smaller steps is the most popular approach, but its appeal differs across regions. In the UK, it’s favored by more than two-thirds of respondents, while in Japan, less than one-third take this approach. Sleep, a cornerstone of effective study, ranks second worldwide. It’s especially valued (by 51%) in Northern and Eastern Europe (including the Balkans and the Caucasus), but less so in Central and South America (29%–36%). Germany stands out, with listening to music edging out getting enough sleep as the top productivity aid (50% vs. 47%). Regular breaks are embraced by learners in the UK, US, Brazil, and Germany, with 46%–51% reporting taking them, but are less common in Japan, South Korea, and China (26%–34%).
Cultural preferences even influence coffee consumption. It’s a favorite pick-me-up, preferred by 37%–41%, in Turkey and all across Northern and Eastern Europe (including the Balkans and the Caucasus), but far less popular among respondents from Nigeria and China (11% and 17%, respectively).
Meanwhile, playing with pets is a go-to strategy in the Americas (10% in the North and 14% in the Central and South), but almost never considered an option in Nigeria, China, South Korea, and the Middle East (1%–4%).
Hobbies learners pursue in their free time100+
Programming Languages and Development
of respondents aged 21–29 report having 3–10 years of general coding experience. This may indicate that people are starting to code earlier than ever before.
Total coding experience (including learning to program and programming as a hobby)
Less than 1 year
1–2 years
3–5 years
6–10 years
11–16 years
16+ years
No coding experience
Professional coding experience
Less than 1 year
1–2 years
3–5 years
6–10 years
11–16 years
16+ years
No professional coding experience
Where learners wrote their first line of code
Integrated development environment (IDE)
Text editor
In-browser code editor
Command-line interface
I’m not sure
Other
Although respondents regard self-paced online tutorials and coding platforms as the top choice for mastering computer science, the IDE remains the most popular option for beginners starting out on their coding journey.
First programming language learned
All answers with less than a 1% share have been merged into “Other”.
Programming languages started or continued in the past 12 months100+
Python dominates both in terms of usage over the past year and ongoing learning, reflecting its widespread applicability and continued growth in popularity. While many learners continue with widely used languages like Java, JavaScript, and SQL, there's also significant interest in newer languages such as Rust and Kotlin.
The data reveals a clear trend of learners expanding their language skills, with a notable focus on foundational languages like Python, Java, and C++, alongside a growing curiosity for emerging technologies.
Programming languages started or continued in the past 12 months100+
Python is in high demand in the United States, with over half of respondents starting or continuing to learn it in the past year. Java learning is most popular in South Korea and India (38%–39%) but much less prevalent in Japan (15%). JavaScript is widely learned in South America and India (40% and 44%, respectively), while TypeScript has seen notable adoption in Germany and France (22%–23%). PHP is far more popular in France than in other regions (16%).
Kotlin is popular in Germany, Spain, South Korea, and the Russian Federation and Belarus (15%–18% in each).
Rust, a functional and system programming language, has gained traction in European regions like France, Germany, Benelux, and Northern Europe (15%–16%).
C++ learning is most popular in India, China, and Ukraine (28%–29%), but much less so in Central and South America, Spain, and Japan (10%–12%). Meanwhile, only 6% of respondents in Central and South America, including Argentina, are learning C, while in India and South Korea, these numbers are four times as high (26%).
Preferred operating systems for development environments100+
Windows
Linux
macOS
Other
Most learners prefer running their code in a local environment, with integrated development environments (IDEs) being the dominant tool. Command-line interfaces and text editors follow as the next most popular choices. Windows is the most widely used operating system for development environments.
Preferred tools for running code100+
Preferred environment for running code100+
Experience with installing and setting up development environments
I’m an experienced user
I’ve set up environments before but might still encounter challenges
I have little experience but have never had issues
I may need guidance or additional resources
I find it challenging and require significant assistance
Other
IDEs / Editors
of all learners reported using an IDE for learning purposes, though the extent of use may vary.
IDEs regularly used for work and learning100+
All answers with less than a 1% share have been merged into “Other”.
Did you know?
Learners who regularly use JetBrains IDEs are 21% more likely to have used an IDE specifically for learning purposes compared to those who do not use JetBrains IDEs. Also, learners who regularly use JetBrains IDEs rate their coding proficiency higher than those who don’t.
Are you a student interested in mastering coding? Get free access to all JetBrains IDEs for personal use at school or at home!
Purposes for using IDEs100+
Personal or side projects
Work
Hobby
Collaborative programming
Other
First IDE used
All answers with less than a 1% share have been merged into “Other”.
Study Routines and Devices
The majority of learners use personal laptops to study computer science and coding. While desktop computers are also commonly used (37% for studying, 36% for coding), smartphones and tablets are less favored, with only a quarter of respondents using smartphones for studying and just 3% for coding. Most learners own their primary study devices, with a smaller percentage relying on devices provided by employers (7%) or educational institutions (3%).
Preferred devices for studying100+
Laptop
Desktop computer
Smartphone
Tablet
Other
Preferred devices for coding100+
Laptop
Desktop computer
Smartphone
Tablet
I don’t write code
Primary study device ownership
I own my study device
My employer provides my study device
I share my study device with my family or housemates
My educational institution provides my study device
Preferred study locations
Convenience of study locations
| Not convenient at all | Fairly inconvenient | Fairly convenient | Very convenient | |
|---|---|---|---|---|
| 1% | 4% | 27% | 68% | Home |
| 2% | 11% | 43% | 44% | Dormitory or student accommodation |
| 2% | 9% | 48% | 41% | Library |
| 1% | 8% | 51% | 40% | School or university campus |
| 1% | 11% | 53% | 35% | Coworking space |
| 1% | 13% | 63% | 22% | Coffee shop |
| 4% | 25% | 50% | 22% | Park or outdoor space |
| 8% | 39% | 38% | 16% | Public transportation (e.g. bus or train) |
Most learners study in the evening, with 58% dedicating 3–16 hours per week to learning computer science. The data reveals that learners would like to spend less time studying in the evenings and at night than they do currently.
Preferred study time
Usual study time100+
Preferred hours per week spent learning computer science
Hours per week spent learning computer science
Preferred learning style
Alone and independently
Combining different studying styles depending on the subject and content
In small peer groups or with a study partner
With a teacher, mentor, or instructor
Undecided
Less than one-third of respondents study systematically, while just over half of respondents don’t follow a concrete study schedule. Key factors impacting the pace of their studies include the workload, deadlines, personal interests, and other personal commitments, all of which play a role in how consistently learners can progress and remain motivated.
Learning pace
I study from time to time; each week I dedicate a different amount of time to learning
I study systematically, learning different topics and trying to dedicate equal time to each one
I study hard for a specific deadline and revert to a more relaxed mode afterwards
Other
Factors influencing learning pace
Respondents answered this question with open-text answers. ChatGPT was used to automate the analysis and sorting of the responses into thematic clusters.
Demographics
Gender
Age group
Gender (by region)
| Prefer not to say | Non-binary, genderqueer, or gender non-conforming | Male | Female | |
|---|---|---|---|---|
| <1% | <1% | 65% | 35% | Russian Federation and Belarus |
| 1% | 1% | 69% | 28% | Argentina |
| 1% | 1% | 71% | 27% | Ukraine |
| 2% | – | 77% | 21% | South Korea |
| <1% | 1% | 80% | 19% | Central and South America |
| <1% | <1% | 81% | 18% | Nigeria |
| 4% | 3% | 75% | 18% | United States |
| 1% | 1% | 81% | 16% | Brazil |
| 4% | 4% | 76% | 16% | Canada |
| 3% | 2% | 79% | 16% | United Kingdom |
| 1% | 1% | 83% | 16% | Middle East, Africa, Central Asia |
| 2% | 2% | 82% | 15% | Spain |
| 1% | 1% | 83% | 15% | Eastern Europe, Balkans, and the Caucasus |
| 1% | 1% | 84% | 14% | Mexico |
| 1% | 1% | 86% | 13% | Benelux and Northern Europe |
| 2% | 2% | 83% | 12% | Japan |
| 3% | 1% | 83% | 12% | France |
| 3% | 1% | 84% | 12% | Rest of Europe |
| 2% | 1% | 86% | 11% | Germany |
| 1% | 2% | 86% | 11% | Türkiye |
| 2% | 1% | 87% | 9% | Other Southeast Asia and Oceania |
| 2% | 1% | 91% | 7% | India |
| 4% | 2% | 90% | 4% | China |
In most regions, the majority of computer science learners are male (80%–90%), with India and China topping that list. On the flip side, higher than average female representation was recorded in the Russian Federation and Belarus, Argentina, and Ukraine.
For France, Germany, and the UK, the figures stand at 11%–16%, which highlights a persistent gender gap in Europe. Non-binary learners make up around 1%–2% in most places, except for the US and Canada.
Marital status
Number of children
of respondents report speaking a different language at home and with friends than the one they use at work. English, Hindi, and Chinese are the top three languages respondents use to speak with their friends and family.
Languages spoken at work100+
All answers with less than a 1% share have been merged into “Other”.
The data shows that English is the dominant language in the workplace, with over two-thirds of respondents using it. Chinese and Japanese are the next most-spoken languages, representing the Asian market. Languages like Hindi, Spanish, and Russian highlight global diversity in tech. Additionally, 8% of respondents use less-common languages not tracked by our survey, indicating even more linguistic diversity in the industry.
Languages spoken with family and friends100+
This question was shown only to respondents who said that they use a different language with friends and family than at work.
All answers with less than a 1% share have been merged into “Other”.
Country / Region
All countries/regions with less than a 1% share have been merged into “Other”.
Mainland China, the United States, India, and Japan combined account for over half of computer science learners worldwide, highlighting the strength of these key global tech hubs.
of respondents were born in a different country or region from where they currently reside, with the Russian Federation, India, and China accounting for one-third of those who have relocated. The migration trend has steadily increased over recent years, with 62% of those who changed countries doing so in the past decade.
Country / Region of birth
This question was only shown to respondents who currently reside in a country or region different from their country or region of birth.
All countries/regions with less than a 1% share have been merged into “Other”.
Methodology
More than 28,500 people took part in the Computer Science Learning Curve Survey 2024.
To ensure a representative sample, we cleaned the data using the method outlined below. The final report is based on responses from 23,991 learners worldwide.
The data was weighted according to several criteria, which are detailed at the end of this section.
Localization
To maximize inclusion and accommodate a diverse range of participants, the survey was available in 10 languages: English, Chinese, French, German, Japanese, Korean, Portuguese, Russian, Spanish, and Turkish.
Sampling-bias reduction
To reduce bias, we weighted the data based on the source of the responses. We prioritized responses from external sources less likely to be biased toward the JetBrains audience, such as paid ads and peer referrals. Each respondent’s source was taken into account individually during the weighting process.
We carried out three weighting stages to ensure a more accurate representation of the global population of computer science learners.
Adjusting for the populations of developers in each region
Before conducting the survey, we did research revealing that the population of STEM students in different regions highly correlates with the number of professional developers in these regions. Based on this finding, we decided to use the proportion of professional developers in each region as an estimate for the proportion of computer science learners.
In the first stage, we assembled the responses from different countries and then applied our estimated distribution of professional developers in each country to weight the data accordingly.
First, we gathered survey responses from ads on social networks across 23 regions, along with responses from peer referrals. Then, we weighted these responses based on our estimates of professional developer populations in each region.
This ensured that the response distribution corresponded to the computer science learner populations in each country.
Adjusting for coding experience and usage of JetBrains IDEs
The second stage involved a more complex process, including calculations based on solving systems of equations.
We used the initially weighted responses to determine the distribution of learners by coding experience level and their use of JetBrains IDEs in each region. These distributions served as constants in our equations.
Next, we added responses from learners who accessed the survey through JetBrains' internal channels, such as our social media accounts and research panel.
Solving the system of linear equations and inequalities
We composed a system of linear equations and inequalities that described:
- The weighting coefficients for the respondents (as a hypothetical example, Fiona from our sample represents, on average, 180 software developers from France).
- The specific values of their responses (for example, Pierre has two years of coding experience and does not use any JetBrains IDEs).
- The necessary ratios among the responses (for example, 22% of learners have 1–2 years of coding experience, and so on).
To solve the system of equations with minimal variance in the weighting coefficients, we applied the dual method of Goldfarb and Idnani (1982, 1983). This approach allowed us to collate the optimal individual weighting coefficients for each of the 23,991 respondents.
Despite these measures, some bias may remain present, as JetBrains’ loyal audience might have been more willing, on average, to complete the survey.
As much as we try to control the survey distribution and apply smart weighting, the communities and the learners ecosystem are constantly evolving, and the possibility of some unexpected data fluctuations cannot be completely eliminated.
In this report, we present a frequency analysis of several open-text questions that received thousands of responses. Owing to the large volume of data, we applied automated processing techniques. To automate response clustering, we used large language models (LLMs), specifically ChatGPT-4o.
Data cleaning
- Translation: To standardize the analysis, responses from different languages were translated into English.
- Completion check: Responses were filtered for relevance and content; irrelevant or off-topic responses were excluded.
- Inappropriate content elimination: Responses containing inappropriate language or offensive content were removed to ensure data integrity.
After the data cleaning procedure, valid responses ranged from 4,000 to 9,000 per question, influenced by the optional nature of some questions and the sensitivity of certain topics.
Response clustering
- ChatGPT-4o was used to analyze and sort the responses into thematic clusters. The analysis was repeated multiple times (typically five or more iterations per question) to validate the consistency of the clusters.
- At each iteration, the clusters were reviewed to ensure that they reflected the core themes of the respondents.
- Testing revealed that 5–8 clusters per question struck the best balance between granularity and generalization, capturing nuanced insights while preserving unique perspectives.
Join the JetBrains Tech Insights Lab
Take part in surveys and UX studies to make JetBrains products easier to use yet even more powerful. For participating in our research, you'll also get the chance to earn rewards.
Thank you for your time!
We hope you found our report useful. Share this report with your friends and colleagues!
Join the conversation using #JBAcademySurveyResults24 to share your thoughts and connect with the dedicated CS learning community on our Discord, LinkedIn, X, and Facebook.